Molten Inorganic Salts as Reaction Medium for Cellulose - ACS

Chapter 4

Downloaded by UNIV OF GUELPH LIBRARY on June 24, 2012 | http://pubs.acs.org Publication Date (Web): February 23, 2010 | doi: 10.1021/bk-2010-1033.ch004

Molten Inorganic Salts as Reaction Medium for Cellulose S. Fischer* and K. Thümmler Institute of wood and plant chemistry, Technische Universität Dresden, Tharandt, Germany *[email protected]

The preparation of cellulose solutions is important for derivatization and blend formation of this natural polymer. Besides solvents like CS2/NaOH and NMMNO*H2O unconventional solvent system can be applied for dissolution of cellulose. This group of solvents includes inorganic molten salts and salt hydrates. Inorganic molten salts can be used as efficient solvents for cellulose in a wide range of degree of polymerization. Furthermore molten salts can be applied as reaction medium for the derivatization of cellulose. For both dissolution and derivatization of cellulose the knowledge of the solution state as well as information about chemical interactions with the solvent system is essential. Beside the specific structure of the molten salt hydrate, the cation and the water content of the melt are the most important factors for the dissolving capability of a molten salt hydrate system.

Introduction The search for new cellulose solvents is still a focus of research. Alternative systems are of special interest for cellulose fiber production and functionalization. Concentrated salt solutions have been known for long time as solvents for cellulose. Intensive investigations have been carried out on swelling and dissolution of cellulose in aqueous zinc chloride. A review about the swelling ability of inorganic salt solutions was published by Warwicker et al. (1). He stated that a lot of salt systems should be able to swell © 2010 American Chemical Society In Cellulose Solvents: For Analysis, Shaping and Chemical Modification; Liebert, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2010.


or dissolve cellulose. But in later publications only three water/salt-systems were described as effective cellulose solvents in more detail: Ca(SCN)2/H2O, LiSCN/ H2O und ZnCl2/H2O. Lukanoff et al. (2) investigated the dissolving ability of the eutectic melt NaSCN/KSCN and of mixtures of this melt with Ca(SCN)2•3H2O. Only the mixture of NaSCN/KSCN with Ca(SCN)2•3H2O or dimethyl sulfoxide was able to dissolve cellulose. In addition the authors described molten LiSCN•2.5H2O as cellulose solvent. The solubility of cellulose with different degrees of polymerization (DP) in Ca(SCN)2•3H2O was discussed by Kuga (3). In the temperature range from 120 °C to 140 °C he observed the solution of the polymer within 40 minutes accompanied by a decrease of DP. There are several papers discussing the formation of addition compounds between cellulose and the dissolving salts. Xu and Chen (4) worked on the dissolution and fiber formation of cellulose in zinc chloride solutions. The regeneration occurred by precipitation of a zinc-cellulose-complex by alcohol. By treatment with water this complex released cellulose II. The formation of the proposed complex was explained using 13C NMR measurements on cellobiose solutions in zinc chloride. Systematic investigations regarding the solubility of cellulose in Ca(SCN)2•3H2O solutions were carried out by Hattori et al. (5, 6). They discussed complex formation between cellulose and Ca(SCN)2 using IR measurements and reported coordination of the Ca2+ ions at O-6 and O-5 of cellulose. These results were also confirmed by 13C and 1H NMR measurements (7).

Solubility of Cellulose in Molten Salt Hydrates A multitude of pure molten salt hydrates as well as salt mixtures were investigated with respect to their interaction with cellulose. As a result it turned out to be reasonable to divide the salt hydrates into groups according to their optical visible effect on cellulose. The classification was done as follows: molten salts which a) dissolve, b) swell, c) decompose cellulose, or which d) have no effect on cellulose. Table 1 gives this classification for the molten salt hydrates investigated. Designation of salt-water-systems: • e.g. LiClO4•3H2O: the hydrate is solid at room temperature marked by the symbol “•” • e.g. ZnCl2+4H2O: the hydrate is liquid at room temperature marked by the symbol “+” • e.g. ZnCl2/MgCl2/H2O: systems of variable compositions marked by the symbol “/” The melts which were able to dissolve cellulose are very different in composition. They vary in cation and anion and in their water content. This results in the question, which factors determine the dissolving ability of the molten salt hydrates. Until now only one rule for salt-water-systems was published. It 92 In Cellulose Solvents: For Analysis, Shaping and Chemical Modification; Liebert, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2010.


says that salts combining small hard cations with soft polarizable anions have the best dissolving power for cellulose. But the good dissolving ability of lithium perchlorate melts observed by us is inconsistent with this statement, and several other hydrated melts containing lithium with different anions do not show any dissolving ability to cellulose (8). After a detailed investigation of the melts the following characteristics, that mainly determine the dissolution power towards cellulose were recognized: • the acidity, • the water content of the melts, • and the properties of the coordination sphere of the cations. (9, 10), respectively. Often these properties influence each other as could be shown by acidity measurements in the system ZnCl2/xH2O depending on the amount of water x (11). A change in the water amount correlates with the acidity, which increases with decreasing x. The acidity parameters of the melting systems ZnCl2/H2O and LiClO4/H2O were investigated using solvatochromic probe molecules. They are comparable to those described for mineral acids. If a hydrated melt should be an effective dissolving agent for cellulose it must have a relative high acidity (12). The dependence of the dissolving power on the water content in different melt systems could be shown for several systems. Molten LiClO4•3H2O for instance is an excellent cellulose solvent. An increase of the water amount up to a composition of LiClO4+4H2O results in a melt which is not able to dissolve the polymer. A significant increase of the swelling grade can be observed by decreasing the water amount of the system LiCl+xH2O from x=5 to x=2. The swelling of cellulose in LiCl+2H2O is strong enough to cause a transformation of the cellulose modification I into modification II. The influence of the structural conditions of the co-ordination sphere of the cations could be explained in Figure 1 by a comparison of the two trihydrates LiClO4•3H2O (solvent) and LiNO3•3H2O (no solvent) (8). Although the salts have the same water content, there are differences in their structures. In the case of the perchlorate the water is completely bridged and bound to the cations. The anions do not affect the co-ordination sphere of the lithium ions. The cleavage of the water bridges results in “free” co-ordination sites. The replacement of water molecules by hydroxyl groups of the cellulose is possible, too. The structure of the nitrate is characterized by water molecules not only bridged between the cations but also at interstitial positions. Because of the bound anions at the cations there is a water deficit at the lithium cations. The nitrate ions can not be replaced by hydroxyl groups because the water from the interstitial positions will prefer to saturate the co-ordination sphere of the cations.

93 In Cellulose Solvents: For Analysis, Shaping and Chemical Modification; Liebert, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2010.

Table 1. Molten salt hydrates and their interaction to cellulose


group

pure melt

melt mixtures

dissolution

ZnCl2+3-4H2O LiClO4•3H2O FeCl3•6H2O

LiClO4•3H2O - ≤25% Mg(ClO4)2/H2O LiClO4•3H2O - ≤10% NaClO4/H2O LiClO4•3H2O - MgCl2•6H2O NaSCN/KSCN (eutectic) – LiSCN•2H2O LiCl/2ZnCl2/H2O

swelling

LiCl+2-5H2O LiNO3•3H2O Na2S•9H2O

LiClO4•3H2O - >25% Mg(ClO4)2/H2O LiClO4•3H2O - >10% NaClO4/H2O

decomposition

Mg(ClO4)2•6H2O MgCl2•6H2O

ZnCl2/MgCl2/H2O

no effect

NaOOCCH3•3H2O

CaCl2•6H2O

Figure 1. Crystal structure of LiClO4•3H2O (left) and LiNO3•3H2O (right)

Structural Changes of Cellulose Dissolved in Molten Salts Hydrates The different interactions between cellulose and the solvent are reflected in the properties of cellulose samples regenerated from the molten salts. The regenerated products were characterized by WAXS, 13C CP/MAS NMR spectroscopy, surfacearea determinations, SEC measurements, solvatochromic and SEM investigations in comparison with the raw material. Using WAXS the transformation of cellulose I into cellulose II could be observed which indicates a strong swelling or dissolution of the polymer. The crystallinity was investigated using solid state NMR spectroscopy (8, 10). Raman spectroscopy has been established too as an effective method to discuss the transformation of cellulose regenerated from molten inorganic salts (13). 94 In Cellulose Solvents: For Analysis, Shaping and Chemical Modification; Liebert, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2010.


The transformation of cellulose I into cellulose II is caused by dissolving cellulose I in different molten inorganic salt hydrates. Whereas, the cellulose regenerated from melts of LiSCN•2.5H2O, LiClO4•3H2O and LiCl/2ZnCl2+8H2O yield by dissolution to a transition into modification II. Beside the X-ray scattering experiments (10) the FT Raman spectra measured after regeneration confirmed the polymorphic changes which are shown in Figure 2. For these regenerated celluloses dissolved by the salt melts mentioned above only the Raman bands characterizing the modification II were detected. There also the intensity maxima which are most typical for following the polymorphic transformation are denoted by crosses. The molar mass distributions determined according a method of Fischer et al. (14) of cellulose regenerated from different molten hydrates are compared with the raw cellulose in Figure 3. It is to be seen that the cellulose regenerated from LiClO4•3H2O exhibits no decrease in the molar mass compared with the raw cellulose, the other samples show a slight shortening of the chain length. The decrease of the molar mass observed here was in the range known for the pre-treatment of cellulose with NaOH-water. Therefore it could be proved that dissolving of cellulose in molten salt hydrates is not due to a drastically reduction of the chain length. Dissolution of cellulose in different melts results in varying morphological properties of the regenerated products. This is indicated by differences in the specific surface-area and the pore size and is clearly recognizable in the SEM pictures. Fiber-like samples similar to the raw cellulose were obtained from the thiocyanate melts, but also lamellar samples (from LiClO4•3H2O) and layered structures (from chlorides) were regenerated (10). Examples for the different morphologies obtained are given in Figure 4.

Interaction between Cellulose and the Molten Salt Hydrate As a very sensitive tool to characterize the state of cellulose in molten salt hydrates NMR spectroscopy can be used. Figure 5 shows the 13C NMR spectra of cellulose dissolved in molten ZnCl2+4H2O, LiClO4•3H2O and LiSCN•2.5H2O. The signals of the carbon atoms C1 – C6 are well resolved and the spectrum is very similar to that of cellulose dissolved in conventional solvents like sodium hydroxide solution (15). A derivatization by the solvent can not be observed. Therefore molten salt hydrates can be attributed to the group of non-derivatizing solvent systems. Due to the different magnetic susceptibilities and melting points of the molten salt hydrates it is difficult to compare the chemical shifts of the carbon nuclei in the different solvents. Nevertheless it can be pointed out that the relative positions of the carbon signals differ for the investigated solvents. Whereas the chemical shifts of the cellulose carbon nuclei are nearly identical in the molten lithium salts, the values for especially C1 to C5 are significantly smaller in the ZnCl2+4H2O melt. We interpret this as an indication for cellulose – solvent interactions of different strength. 95 In Cellulose Solvents: For Analysis, Shaping and Chemical Modification; Liebert, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2010.


Figure 2. FT Raman spectra of cellulose II regenerated from different molten salts hydrates in the range 1500 - 150 cm–1 Further information about the interaction between cellulose and molten inorganic salts can be acquired by 7Li NMR investigations (16) and Raman measurements (17). Röder (18) investigated the structure of cellulose dissolved in molten LiClO4•3H2O by static laser-light-scattering. The investigation of cellulose solutions in molten LiClO4•3H2O by static light-scattering measurements at high temperature showed that cellulose molecules form aggregates. The solvent is not able to destroy the intermolecular hydrogen bonds between cellulose chains completely. Single cellulose molecules could not be found. A comparison of the measured form factor function with theoretical models showed that the shape of the particles is between a globular and a homogenous branched structure. The increase of the function in the Kratky plot can be explained with a second component – small particles or arms of fringed micelles. The analysis indicates for cellulose in molten salts comparable structures like for solutions of cellulose in N-methylmorpholine-N-oxide-monohydrate (19).

Reactions of Cellulose in Molten Inorganic Salt Hydrates Molten inorganic salt hydrates as so-called non-derivatizing solvents were not only established as very efficient solvents for cellulose, they can also be applied as a medium for the chemical functionalization of the polymer. Using molten inorganic salt hydrates as new reaction media etherification (e.g. carboxymethylation) and esterification (e.g. acetylation) of cellulose are possible and will be discussed here. 96 In Cellulose Solvents: For Analysis, Shaping and Chemical Modification; Liebert, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2010.


Figure 3. Molar mass distribution of the nitrates of the cellulose samples regenerated from P2: LiClO4•3H2O, Z3: ZnCl2+4H2O, T4: NaSCN/KSCN/LiSCN/H2O, C: raw cellulose

Figure 4. SEM pictures of celluloses regenerated from NaSCN/KSCN/LiSCN/H2O (left) and LiClO4•3H2O (right) Carboxymethylation Because of the fact that the solvent LiClO4•3H2O was extensively studied (8, 9) preliminary experiments concerning its use as reaction medium were carried out (20). The results were very promising and therefore systematic studies regarding carboxymethylation of cellulose in different inorganic molten salts were started. The method and the preparation procedure were described (21). The homogeneous carboxymethylation of cellulose in molten LiClO4•3H2O is possible using sodium monochloracetate in the presence of NaOH. 97 In Cellulose Solvents: For Analysis, Shaping and Chemical Modification; Liebert, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2010.


Figure 5. 13C NMR spectra of cellulose dissolved in molten a) ZnCl2+4H2O at 65 °C, b) LiClO4•3H2O at 110 °C, c) LiSCN•2H2O at 130 °C A remarkable finding is that polymers with DS values as high as 2 can be prepared within a short reaction time (4 hours) applying a one-step synthesis. Up to now this was only possible by conversion of cellulose in DMA/LiCl or using the so-called induced phase separation (22, 23) which starts from organo-soluble hydrolytically unstable intermediates. In the case of homogeneous carboxymethylation or reaction in slurry the highest DS reachable in a one step procedure is about 1.3. Selected samples were characterized by means of 1H-NMR spectroscopy after chain degradation with a mixture of D2SO4/D2O (24, 25). A distribution of substituents on the level of the AGU in the order C-6 > C-2 ≈ C-3 was discovered and the investigations showed that a complete substitution at position O-6 is possible. Further investigations showed that carboxymethylation is also possible in the system LiCl-H2O starting from swollen cellulose. The DS of CMC prepared in molten LiCl+xH2O, i.e. in a heterogeneous reaction, is generally lower than the DS of the samples obtained from dissolved cellulose in molten LiClO4•3H2O. The DS also depends on the molar ratio AGU:sodium monochloracetate:NaOH and the DS decreases with increasing water content of the molten salt hydrate. The samples have a similar distribution of carboxymethyl groups like CMCs prepared under homogeneous conditions. The results show a distribution of substituents on the level of the AGU in the order C-6 > C-3 ≈ C-2. That means, there is also a difference between carboxymethylation in this swelling medium and conventionally produced CMCs.


Comprehensively it can be said that a etherification like carboxymethylation of cellulose in hydrated melts yields products of a high degree of substitution of up to 1.96 in an one step synthesis within a short reaction time. Consequently, molten LiClO4•3H2O is an efficient solvent for carboxymethylation of cellulose yielding highly functionalized polymers. Moreover, a specific influence on the distribution of functional groups on the level of the AGU is accessible.


Acetylation Several molten inorganic salt hydrates were applied as media for the acetylation of cellulose. The formation of cellulose acetate depended on the used molten salt hydrate itself as well as on the water content of the melt (26). However, the only melt in which the acetylation was successful is the eutectic mixture of NaSCN and KSCN with addition of 10% LiSCN•2H2O. The most important condition for the success of the reaction was the minimization of the water content of the melt. The acetylation was carried out at a temperature of 130 °C with a high excess of acetanhydride (50–100%). During a short reaction time (0.5–3 h) cellulose acetate with a DS in a range between 1 and 2.5 were obtained. The samples were characterized by IR-spectroscopy. For a detailed characterization X-ray and NMR-measurements were carried out. By these investigations it could be shown that the acetylation in molten thiocyanate leads to the formation of amorphous cellulose acetate (27). The DS obtained for the cellulose acetates depends on the reaction time and on the molar ratio between AGU and acetanhydride. For future work we will extend the use of molten salt hydrates to other applications. So it is possible to use the acidic properties of the molten inorganic hydrates for the cleavage of functional groups. For instance cellulose triacetate deacetylation can be carried out in molten ZnCl2+4H2O. After a reaction time of 21h a DS of 1.81 and PDS (C6) of 0.53 were obtained (10). Furthermore molten ZnCl2+4H2O or LiClO4•3H2O can be applied as medium for deprotection of triphenylmethyl cellulose. A complete deprotection occurs in the comparatively short reaction time of 3 to 5 hours (28).

Blend Formation Use of molten salts as medium for the blending of cellulose with synthetic polymers was studied. It was shown that formation of cellulose-polyacrylonitrileblends is possible by dissolving a mixture of PAN and cellulose in molten thiocyanates simultaneously. KSCN/NaSCN/LiSCN•2H2O is the first choice for that process because it can dissolve both the naturally and a synthetically polymer completely. Regeneration of the blend from the polymer solution can be simply achieved by cooling the melt to room temperature and treating the polymer/salt mixture with water for a complete removal of the thiocyanates. A homogeneous polymer blend was obtained by this simple procedure. 99 In Cellulose Solvents: For Analysis, Shaping and Chemical Modification; Liebert, T., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2010.


The thermal properties of the blend were determined. These blends show similar behavior, as described for material made from DMF/N2O4 solutions. Examining the structure of the blends by means of NMR spectroscopy did not provide direct information about interactions between the polymers. Nevertheless, the change in the intensity of nitrile signals is a first hint for interactions between the polymer components initiated via this moiety. In principle, the use of molten salts allows a broad variety of crystallization conditions. It can be distinguished between very quickly crystallizing systems on one hand and systems with extreme tendency towards hypothermia on other hand. The variation of coagulation conditions and the resulting properties of blends are important for tailored preparation of polymer blends and will be the topic of further investigations.

Conclusion Molten salt hydrates are effective and efficient media for cellulose dissolution. During the investigations regarding the solubility of cellulose in molten salts new solvents were found (LiClO4•3H2O). First time factors which determining solution ability of cellulose in molten inorganic salt hydrates was found. This knowledge is a basis for discovering further molten hydrates as solvents for cellulose. The structural change of cellulose after dissolution depends on the respective hydrated molten salt. Therefore it should be possible to adjust specific cellulose structures by choosing a certain salt melt for cellulose dissolution and regeneration. For cellulose carboxymethylation and acetylation it was shown exemplary, that molten salt hydrates are efficient solvents for cellulose derivatization yielding highly functionalized polymers. The results are a basis for further investigations using molten inorganic salts as a reaction medium for cellulose, which will be extended to other polysaccharides.

References 1. 2. 3. 4. 5. 6. 7. 8. 9.

Warwicker J. O.; Jeffries R.; Colbran I. ; Robinson R. N. Shirley Institute Pamphlet. No. 93, Manchester, 1966. Lukanoff, B.; Schleicher, H.; Philipp, B. Cell. Chem. Technol. 1983, 17, 593–599. Kuga, S. J. Colloid Interface Sci. 1980, 77, 413–418. Xu, Q.; Chen, L. F. Text. Technol. Int. 1996, 40, 19–21. Hattori, M.; Shimaya, Y.; Saito, M. Polymer J. 1998a, 30, 37–42. Hattori, M.; Shimaya, Y.; Saito, M. Polymer J. 1998b, 30, 43–48. Hattori, M.; Shimaya, Y.; Saito, M. Polymer J. 1998c, 30, 49–5. Fischer, S.; Voigt, W.; Fischer, K. Cellulose 1999a, 6, 213–219. Fischer, S.; Leipner, H.; Brendler, E; Voigt, W.; Fischer, K. ACS Symposium Series 1999b, 737, 143–150.



10. Leipner, H.; Fischer, S.; Brendler, E.; Voigt, W. Macromol. Chem. Phys. 2000, 201, 2041–2049. 11. Leipner H.Dissertation A, TU Bergakademie Freiberg, Germany, 2002. 12. Fischer, S.; Voigt, W.; Fischer, K.; Spange, S.; Vilsmeier, E. Molten Salt Forum 1998, 5–6, 477–480. 13. Schenzel, K.; Fischer, S. Cellulose 2001, 8, 49–57. 14. Fischer, K.; Schmidt, I.; Hintze, H. Papier 1994, 48, 769–774. 15. Nehls, I.; Wagenknecht, W.; Philipp, B.; Stscherbina, D. Prog. Polym. Sci. 1994, 78, 1929–1979. 16. Brendler, E.; Fischer, S.; Leipner, H. Cellulose 2002, 8, 283–288. 17. Fischer, S.; Leipner, H.; Thümmler, K.; Brendler, E.; Peters, J. Cellulose 2003, 10, 227–236. 18. Röder, T. Dissertation A, TU Dresden, Germany, 1998. 19. Röder, T.; Morgenstern, B. Polymer 1999, 40, 4143–4147. 20. Heinze, Th.; Pfeiffer, K. Angew. Makromol. Chem. 1999, 266, 37–45. 21. Fischer, S.; Thümmler, K.; Pfeiffer, K.; Liebert, T.; Heinze, T. Cellulose 2002, 9, 293–300. 22. Heinze, Th.; Erler, U.; Nehls, I.; Klemm, D. Angew. Makromol. Chem. 1994, 215, 93–106. 23. Liebert, T.; Heinze, Th. ACS Symposium Series 1998, 688, 61–72. 24. Gronski, W.; Hellmann, G. Papier 1987, 41, 668–672. 25. Baar, A.; Kulicke, W-M.; Szablikowski, K.; Kiesewetter, R. Macromol. Chem. Phys. 1994, 195, 1483–1492. 26. Thümmler K.; Fischer S.; Peters J.; Liebert T.; Heinze Th. Cellulose, 2009, in press. 27. Fischer S. Habilitationsschrift, TU Bergakademie Freiberg, Germany, 2004. 28. Fischer, S.; Leipner, H.; Liebert, T.; Heinze, Th. Polym. Bull. 2001, 45, 517–521.


Molten Inorganic Salts as Reaction Medium for Cellulose - ACS

Recommend Documents